Bug 128895 - planet.g.o. republishes same stuff over and over again
Bug#: 128895 Product:  Website www.gentoo.org Version: n/a Platform: All
OS/Version: Linux Status: RESOLVED Severity: major Priority: P2
Resolution: FIXED Assigned To: dsd@gentoo.org Reported By: jakub@gentoo.org
Component: Planet
URL: 
Summary: planet.g.o. republishes same stuff over and over again
Keywords:  
Status Whiteboard: 
Opened: 2006-04-05 04:38 0000
Description:   Opened: 2006-04-05 04:38 0000
Planet has been broken for a while. Republished that hanno's amarok + moodbar
thing many times, republished dsd's complete blog and today it just republished
zaheerm's complete blog.

------- Comment #1 From Daniel Drake 2006-04-06 17:42:56 0000 -------
Nothing planet can do about this. My blog was reposted when I upgraded
wordpress, wordpress changed all the timestamps. Same with zaheer I guess.
Hanno's blog writes incorrect timestamps every time he writes a new post, even 
in category unpublished on the planet. Hanno is aware, hopefully he has it
fixed.

------- Comment #2 From Jakub Moc (RETIRED) 2006-04-24 03:38:23 0000 -------
OK... Today: complete nixphoeni's blog (universe looks even worse, even more
outdated stuff) republished; and sebastian's eZ components appear yet again.
Meanwhile, the amarok thing has been republished a bunch of times.

If this can't be fixed on p.g.o level, then remove the broken blogs from
aggregation until they are fixed. Ignoring the breakage makes Gentoo look
pretty lame. :/

------- Comment #3 From Daniel Drake 2006-04-25 14:52:29 0000 -------
Created an attachment (id=85489) [details]
Hanno's RSS feed, 25th april

Hanno,

Please fix your blog. I am attaching a current copy of the RSS feed for future
reference.

If the dates change again then I'll temporarily remove you from the planet
until it is fixed. Sorry to sound so harsh, just it does damage the flow of the
planet...

------- Comment #4 From Daniel Drake 2006-04-25 14:56:19 0000 -------
Created an attachment (id=85490) [details]
Joe Sapp's feed, April 25th

Joe Sapp (nixphoeni)'s weblog has only been republished once to my knowledge..
probably due to a blog software upgrade? Anyway, here's a current copy of the
RSS feed, if it happens again we can investigate further.

------- Comment #5 From Daniel Drake 2006-04-25 15:00:48 0000 -------
Created an attachment (id=85492) [details]
Sebastian Bergmann's feed, 25th april

Sebastian,

Planet has been repeatedly republishing your weblog on the site. This usually
happens when your buggy blog software mistakenly modifies the date of every
post in your feed, or if it makes some other large unneeded change.

Here is a copy of your feed from today. If it happens again, we'll have a basis
for comparison.

------- Comment #6 From Sebastian Bergmann (RETIRED) 2006-04-26 01:06:21 0000 -------
AFAIK, Serendipity, which is the blog software I am using, bumps the RSS
timestamp when an entry is edited after its initial publication. I would not
call this behaviour a bug, but a feature.

------- Comment #7 From Daniel Drake 2006-04-26 02:07:30 0000 -------
The post which springs to mind is the "EZ Components now available in Portage"
one. Have you been editing that one?

------- Comment #8 From Sebastian Bergmann (RETIRED) 2006-04-26 02:18:12 0000 -------
Okay, I just checked with Garvin Hicking (lead developer of Serendipity).

The current behaviour is this: Serendipity's feeds contains all entries since
If-Last-Modified, and not only the last 15 (which would be sent for an
Unconditional GET request). Planet does not handle this correctly, hence the
"republishing".

The issue is fixed in Serendipity's SVN TRUNK (which will become Serendipity
1.1): "Added new configuration option to that lets you toggle whether to comply
with RFC2616 on Conditional GET.".

------- Comment #9 From Sebastian Bergmann (RETIRED) 2006-04-26 02:27:29 0000 -------
Yes, I edited that posting, AFAICR.

------- Comment #10 From Joe Sapp 2006-04-26 22:27:16 0000 -------
(In reply to comment #4)
> Joe Sapp (nixphoeni)'s weblog has only been republished once to my knowledge..
> probably due to a blog software upgrade?

Sorry about that.  I was screwing with the RSS flavour I wrote for blosxom, so
the content of the feed changed about 3 times.  Planet wasn't catching when I'd
post something new, so I tried to fix it.

------- Comment #11 From Hanno Boeck 2006-04-27 03:18:16 0000 -------
I've installed the svn-version mentioned by sebastian in #8 and activated the
option, so my blog should be okay now.

------- Comment #12 From Brian Harring 2006-05-01 00:32:00 0000 -------
@hanno: it ain't, repeat as of 05/01/06 07:31 utc.

------- Comment #13 From Daniel Drake 2006-05-03 12:18:41 0000 -------
Ok, I just discovered that we were running an ancient version of Planet (the
"nightly snapshot" on planetplanet.org was like years old). I just upgraded us
to the latest nightly as of today. Fingers crossed that this has better old
article detection, since that TODO item disappeared...

(I think the republishing of Hasan Khalil's blog is a side-effect of the
upgrade.)

------- Comment #14 From Daniel Drake 2006-05-30 16:19:47 0000 -------
Marking as fixed for now, problem doesn't seem to have reappeared